corner-left

Marketing AI: How Not To Make An A.R.S.E. Of It

corner-right

MachineLearningMarketing

There isn’t a single marketer, agency or tech company out there who isn’t talking about the opportunities, threats and challenges we face with the growth of artificial intelligence in our lives and in our industry.

The whole of CES was buzzing with AI promises, debates, and showcases.

The only thing is, I don’t think we are really talking about artificial intelligence.

When I was a teenager, I devoured books in general, and developed a fascination with science fiction in particular. My favorites were books by Heinlein, Herbert, Clarke and Asimov. The last author is, of course, famous for penning the three laws of robotics back in 1942, and adding a forth (Law Zero, listed below first) in the early 1960s:

  • A robot may not harm humanity, or, by inaction, allow humanity to come to harm.
  • A robot may not injure a human being or, through inaction, allow a human being to come to harm.
  • A robot must obey the orders given it by human beings except where such orders would conflict with the First Law.
  • A robot must protect its own existence as long as such protection does not conflict with the First or Second Laws.

A robot can only exist with some form of artificial intelligence. In fact, you could argue that any kind of algorithm that is programmed to make choices is a form of AI. This means that programmatic buying, for instance, is AI. But equally, your Roomba vacuum cleaner, Amazon’s Alexa, the lane assist in your car and that bot that is answering your chats on a website is a form of AI.

But here is the thing. As smart as all of these systems are, I don’t think they are artificial intelligence in the way Asimov and other science fiction writers envisioned it. I believe they are better described by what I would call automated responsive software engines. And woohoo, that spells ARSE!

Because let’s face it: what we are talking about to date are not smart, autonomous and intelligent decision-making systems that replicate human decision-making. Rather, these are incredibly sophisticated automated decision tree calculators. If (X) occurs, check (a), (b) and (c), then determine the most favorable outcome.

Now I am very much in favor of technical advancement, and this is not a piece to denounce or belittle the tech of today and the future. But I do believe that this kind of AI, like robotics, require a set of laws to ensure it doesn’t do harm. So I am proposing that AI systems that govern marketing- and/or advertising-related decision-making adhere to the following laws:

  • Marketing and Advertising AI must keep a permanent and accessible record of the decisions, parameters and alternatives it considered, and inputs that influenced its decision-making process.
  • Marketing and Advertising AI must be driven to deliver the best result possible within the parameters available.
  • Marketing and Advertising AI must protect its decision-making from manual overrides that would lead to a sub-optimal outcome.

That is my first attempt. I am sure that these are very much open for improvement. I am also very aware that the chances of these actually being adopted and put in place are probably zero. And I calculated that without the help of an algorithm…

[Maarten is a featured contributor to MediaPost, this article was originally published here]

Leave a Comment